A New version of the NVIDIA app The tricks are developed with a beautiful feature. With big news NVIDIA app version 11.0.3.218 Does your computer run locally?
The first feature includes the ability to advance base resolution for DLSS with an advanced setting with single digits granularity. Inside the app you can set a base resolution from where the DLSS scales on each game, the game developer has selected any setting over -riding in which the game developer selected for DLSS settings.
“Base Resolution” means that the GPU’s 3D pipeline adds pixels to produce the ultimate image of high resolution. For example, for the last advanced resolution of 2,160 pixels by 4K or 3,840, the base resolution can be 1,920 or 1080p to increase performance mode, while the quality mode will usually be 1440p base resolution or 2,560 1,440.
Between 33 % and 100 % of the final advanced resolution can be selected anywhere. If 100 % sounds such as it does not mean -100 % base resolution does not mean at all? -This DLAA or deep learning anti-elastic implementation, which is a more efficient and fast anti-elephant routine that traditional methods such as MSAA or multi-sampling AA.
This is a beautifully useful feature to tonna your own DLSS methods. But this is the project G Assist which can make more change for more and more gamers. It is an AI assistant based on a small language model that runs locally on your computer.
For now, it is only compatible with the RTX 30 series and the desktop GPUs. Nvidia says laptop GPU support will be added later.
However, NVIDIA says GS estate can help users “improve sports and system settings, improve frame rate and other important performance statistics, from controlling selected prefrilles’ settings such as lighting, through basic sound or text commands.”
It is unclear how flexible the natural language interface will be. In an ideal world, you will be able to say something, “Hey, half -life 2 RTX is running badly, can you make it a bit smooth without affecting the quality of the image,” and then after seeing the results, “It’s not bad but the texture can make you look a little, but you may look like it,” but you can make it look like, but you can make it look faster, but you can make it look like, ” “
Project G Assist is available as a separate download from the NVIDIA app’s home tab in the “Discovery” section. Note, it will be visible only when you have a GPU compatible.
Anne Wadia says G. Assist Lama uses a model -based model, which has eight billion parameters, which is relatively small compared to large language models like Chat GPT4, with about 1.8 trillion parameters. Small size of the model means that it can run locally on the Gaming GPU.
For now, we do not know how much Vram G-Assist is used. Given that the GPUS video at the bottom of the NVIDIA is slightly lower than that of video memory, as it is, it is a bit concerned. But then maybe, maybe, if such features become more important, NVIDIA will be involved in allocating VRAM, as Apple has increased its Max base from 8GB to 16GB to support AI features and reduce the Intel’s luminous lake chip to reduce Microsoft’s Copylot.
Well, I hope to expect, okay?!?
